The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to launch and manage the codeplain renderer process.
Evidence: Launches the renderer using nohup codeplain <module>.plain --headless <other-flags> > /dev/null 2>&1 &.
Evidence: Uses tail, grep, and wc to monitor log growth and content incrementally.
Evidence: Manages the process lifecycle using ps for monitoring and kill -INT for termination.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it ingests and processes external data to drive its decision-making logic.
Ingestion points: The skill reads from codeplain.log, .memory/conformance_test_memory/<id>.json, and project-defined test script outputs in SKILL.md.
Boundary markers: No delimiters or boundary markers are used to isolate untrusted external content from the agent's internal reasoning context.
Capability inventory: The skill can terminate processes, launch shell commands, and delegate tasks to other skills based on the content of the logs it reads.
Sanitization: No sanitization or validation of the log content is performed before it is processed by the agent's classifier logic.

run-codeplain